Home

Gliomas Pituitary adenomas Meningiomas Craniopharyngiomas Ependymomas Schwannomas Pinealomas MRI in Brain Tumors

Neuroanesthesia

Contact Us

Semiotics of proton MR spectroscopy in brain tumors
Proton MR spectroscopic imaging (1H-MRSI) combines the spatial localization capabilities of MR imaging with the biochemical information of 1H-MR spectroscopy. A multitude of spectroscopic studies has been published in the literature in the last 20 years. 1H-MRSI in brain tumors is feasible, reasonable, and a valuable clinical tool. These studies have consistently shown that the Cho signal is elevated in all tumor types because of altered membrane metabolism. The Cho (3.24 ppm) signal increases with cellular density and, according to some authors, also correlates with indices of cell proliferation (Ki-67). The signal intensity of N-acetyl-aspartate (NAA, 2.02 ppm), a marker of neurons and their processes, decreases with tumor infiltration and substitution of normal neural and glial cells. Changes in the creatine (Cr, 3.02 ppm) signal may vary with tumor type: there is often a mild increase in LG astrocytomas, followed by progressive depletion with increasing anaplasia. Unfortunately, at the spatial and spectral resolutions available with clinical studies, no metabolite is tumor type-specific.
Elevation of the Cho signal has been recognized as an important surrogate marker of tumor progression and response to therapy. Precursors and catabolites of phospholipid metabolism are altered in tumors. The total choline peak detected by clinical 1H-MRS is the sum of the signals of free Cho and Cho-containing compounds: phosphocholine (PC), and glycerol 3-phosphocholine (GPC), with perhaps small contributions also from phosphoethanolamine (PE), and glycerol 3-phosphoethanolamine (GPE). Elevation of PC and PE, intracellular phosphomonoesters (PME), suggests enhanced cell membrane synthesis during cellular growth, while elevation of GPC and GPE, phosphodiesters (PDE), is due to abnormal rates of membrane synthesis, catabolism, and metabolic turnover. Oncogene expression and malignant transformation have common endpoints in choline phospholipid metabolism. They often determine a shift in PC and GPC that results in increasing PC/GPC ratio and total choline signal. Experimental data suggest that increased PC/GPC is primarily related to malignant degeneration; conversely, reduced PC/GPC is related to growth arrest. Ex vivo studies have also shown that PC elevation is likely related to malignancy, since it is found at a twofold greater level in HGG compared to LGG and normal tissue. Therefore, MRS appears to be more sensitive to up-regulation of the anabolic pathway than acceleration of the catabolic pathways.
In vivo 1H-MRSI studies have shown that Cho elevation is highest in the center of a solid, non-necrotic glioma with a Cho/NAA gradient declining toward the periphery of the tumor. The slope of this gradient is frequently shallow toward the cortex, while it is frequently steeper toward the white matter. One study in 18 glioma patients demonstrated a significant linear correlation between normalized choline (nCho) signal and cell density, not with proliferative index. The same study found a significant inverse linear correlation between cell density and the apparent diffusion coefficient (ADC) measured by diffusion MR. The Cho signal is consistently low in areas of necrosis.
In a recent prospective study specifically devoted to the characterization of diffuse WHO grade II gliomas, a significant correlation between Ki-67 labeling index and single-voxel MRS data was found. It was shown that increasing Cho/NAA and Cho/Cr ratios correlated with three classes of Ki-67 indices. The presence of lactate was found only in LGG with Ki-67 > 4%; the presence of lipids was found only in LGG with Ki-67 > 8%. Although the Ki-67 labeling index is not formally considered as a prognostic marker in the WHO classification, an index over 4% is predictive of a more aggressive clinical behavior. This study provides evidence that 1H-MRS may be a reliable tool for predicting the proliferative activity of WHO grade II gliomas. Observation of increasing Cho/NAA and Cho/Cr
ratios in consecutive 1H-MRSI exams is highly suggestive of LGG transformation (Figure 1). Tedeschi et al. have shown that serial 1H-MRSI examinations can detect transformation of LGG and distinguish between progressive and stable gliomas. In this longitudinal study of 27 patients, the authors demonstrated an interval increase in Cho signal of greater than 45% in patients with progressive disease, while interval Cho increase was 20% or less in stable patients.
The Cr signal is the sum of Cr and phosphocreatine (PCr). These two molecules play an important role in short-term energy storage, with the Cr kinase reaction generating ATP. Cr is synthesized in the liver and used in the brain (slightly higher in gray than in white matter), in the kidneys, heart, and skeletal muscles. It is absent in all other tissues. In brain tumors Cr signal may change according to tumor type and glioma grade. Cr is virtually absent in lymphomas and metastases, very low in meningiomas and oligodendrogliomas. In LG astrocytic tumors and in gliomatosis cerebri Cr signal may be elevated. Cr may be as elevated as the Cho signal in grade II glioma infiltrating the thalamus and other gray matter structures. However, Cr signal may drop substantially as signs of anaplasia increase. The Cho/Cr ratio is often much greater than unity in primary de novo and secondary GBM.
Mobile lipids (1.40 and 0.9 ppm) are a characteristic feature of GBM, metastases, lymphomas, and abscesses. Abnormal signals resonating at 1.4 and 0.9 ppm have been associated with lipid droplets in areas of extracellular necrosis. Primary cerebral lymphoma in immunocompetent patients mimics the infiltrative behavior of glial neoplasms. The diagnosis of lymphoma is difficult with conventional MRI, and it has important diagnostic and therapeutic implications: surgery should be avoided, chemo- and radiotherapy are the treatments of choice. A spectrum with high Cho, absent Cr and NAA and elevated lipid is the hallmark of cerebral lymphomas; however, it is not the rule in all patients with CNS lymphoma. Large amount of lipids can also be found in areas that have been treated with
radiotherapy and have evolved into delayed radiation necrosis.
The hypothesis that accumulation of lactate (1.34 ppm) may correlate with higher tumor grade was investigated in the early days of MR spectroscopy. As originally described by Warburg, neoplastic cells may develop bioenergetic aberrations, such as elevated anaerobic glycolysis. This behavior is mainly characteristic of HGG that have lost aerobic cell respiration capability. Anaerobic glycolysis is less efficient and leads to increased production of lactate. 1H-MRS studies have shown that lactate accumulation occurs in about one third of gliomas. It is found more frequently in HGG, but its presence is not a reliable indicator of tumor grade. Lactate can be found occasionally in grade II gliomas. Lactate accumulation may occur in areas of tumor with low rCBV and slow flow (low mean transit time [rMTT]). The detection of lactate accumulation in an LGG may be a sign of anaerobic glycolysis occurring in proliferating cells, or due to relative hypoxia. This lactate sign may precede increased vascular density and angiogenesis that is detected by MR perfusion on CBV maps. Thus lactate accumulation in LGG may be a transient phenomenon that will eventually disappear if induction of angiogenesis facilitates lactate clearance through the venous drainage. Increased lactate within necrotic pseudocysts and in areas of the tumor where venoular outflow is obstructed is also a relatively frequent finding. Lactate may also be detected in the necrotic areas of GBM and metastases.
The NAA signal is a surrogate marker of neurons and its processes in gray matter and of axons in white matter. Tumor infiltration is associated with displacement and destruction of normal tissue; thus, NAA is markedly reduced in the core of most solid brain tumors. In the core of meningioma, metastasis and lymphoma absence of the NAA peak on the spectrum is a frequent finding since these tumor types do not contain NAA. In gliomas, which grow by infiltration, the detection of different amounts of NAA signal in the spectra may indicate residual neurons or axons. Whether this residual neural tissue is viable cannot be determined with MRS. Other imaging methods could be used to address this question. Another explanation for residual NAA signals within gliomas is partial volume average with adjacent normal tissue due to the limited spatial resolution of MRS.
The role of myo-inositol (mI) in LGG remains controversial. In a very recent study it was shown that low-grade oligodendroglioma may have higher mI/Cr ratio than astrocytoma. However, previous studies have shown abnormally elevated mI in LG astrocytomas as well. Elevation of mI and Cr has been reported also in gliomatosis cerebri. Elevation of mI and Cr in LGG is a relatively common finding; however, it appears to be more reflective of the pattern of tumor infiltration in the gray matter rather than of the tumor type.
Alanine resonates at 1.47 ppm and is occasionally found in the spectrum of meningioma and abscesses. Alanine is a J-coupled, doublet peak (similar to lactate) and it will be inverted at TE values between ~135 and 144 ms (1/J), which can help with assignment. It
may overlap with lactate and form an apparent “triplet” peak that will be inverted at TE = 135–144 and positive at TE = 270–288 ms. In a recent study, alanine was found in 10 of 31 meningiomas (32%). Alanine and lactate were overlapped in 8 of the 10 cases. Nine of the meningiomas were benign, one was atypical. However, the reported occurrence of alanine varied greatly according to different studies, from 0/6 to 21/23. One study showed that this variance may arise from two technical aspects: failure to recognize the overlapping peak of alanine and lactate, and the small voxel size. Alanine is produced during transamination between alanine and α-ketoglutarate, instead of glycolysis. Alanine is less commonly detected in malignant meningiomas probably because glycolysis plays a greater role for the large energy supply in these tumors.
There are several indications for an MRS study: differentiation of neoplastic versus non-neoplastic lesions, definition of brain tumor type, grading of glioma, assessing response to therapy, and differentiation of recurrent tumor versus delayed radiation necrosis. In the evaluation of brain tumors, multivoxel (2D and 3D 1H-MRSI) is superior to single voxel methods for at least three reasons. 1H-MRSI allows the evaluation of spatial heterogeneity and the identification of the macroscopic boundary of a mass, and it is less prone to partial volume artifacts. Spectroscopic imaging is valuable in characterizing areas of T2-weighted signal hyperintensity: it may delineate areas with high cellular density; it may distinguish areas with prevalent vasogenic edema from areas with neoplastic invasion or necrosis. The choice of the echo time (TE) is also important. Historically, the majority of 1H-MRSI studies have been performed with intermediate (136 ms) or long (272 ms) TE. The longer TE usually provides simpler, cleaner spectra that are easy to interpret, with flatter baselines and less lipid contamination from the scalp. Short TE (30ms) is recommended when evaluation of metabolites with shorter T2 values such as myo-inositol, glutamate/glutamine, and lipids are needed.
Interpretation of 1H-MRSI studies will also be more robust if spectra are acquired with high spatial resolution.
Five frequently asked questions
Is it a tumor?
This is one of the most frequently asked questions (FAQ) to radiologists. Spectroscopy is valuable in cases when conventional MR imaging and the clinical history are ambiguous. When the differential diagnosis includes stroke, focal cortical dysplasia or herpes encephalitis and neoplasm, the finding of an elevated Cho peak makes the diagnosis of tumor more likely. There are caveats as well, and the radiologist must be well aware that some non-neoplastic focal lesions may show elevated Cho. For example, an acute giant demyelinating plaque could mimic an HGG on both MRI and 1H-MRS. Acute demyelinating lesions may show elevated Cho and decreased NAA signal. Successful classification of resonance profiles in 66 gliomas and 5 acute demyelinating lesions using a leave-one-out linear discriminant analysis has been reported. Cho elevation in the 1H-MR spectrum has been also been reported in Erdheim–Chester disease (ECD), a rare non-Langherans histiocytosis that presents in adults. ECD is considered an inflammatory process that is believed to originate from mononuclear phagocytes which proliferate and infiltrate multiple organs, including the brain. Infiltrative nodules or masses in the pons, cerebellum, and hypothalamic-pituitary axis may simulate an infiltrating enhancing tumor on post-contrast MRI.
Is it a GBM, metastasis, or an abscess?
The second FAQ is the differential diagnosis of a ring enhancing mass. The best strategy is to use a multivoxel PRESS sequence with intermediate TE to look for elevation of Cho in the enhancing rim and in the peri-lesional T2 hyperintensity. If Cho is elevated in both areas, a likely diagnosis of GBM may be suggested. Elevation of Cho in the enhancing rim but not in the surrounding tissue would suggest the diagnosis of metastasis. In spectra derived from the necrotic/cystic core of the mass, accumulation of lipids or lactate without elevated Cho is not a specific finding; thus, the acquisition of an additional single voxel spectrum with short TE would be helpful to detect the presence of other minor peaks beyond lactate: succinate (2.42 ppm), acetate (1.9 ppm), or amino acids such as leucine (3.6 ppm), alanine (1.5 ppm), and valine (0.9 ppm). The detection of few or all of these peptides and amino acids confirms the diagnosis of a pyogenic abscess.
What’s the grade of this glioma?
Whether 1H-MRSI is useful for grading of gliomas or not remains a controversial issue. There is a body of evidence in the literature that both Cho/NAA and Cho/Cr increase with cellular density and mitotic index. However, in the individual cases it may be difficult to assign a grade to a mass on the basis of 1H-MRS alone. It is therefore useful to review changes in metabolic profile occurring during the malignant transformation from diffuse to anaplastic astrocytoma: the NAA signal falls to the baseline while the Cho signal increases with higher cell density and proliferation. The Cho/NAA ratio is likely the most sensitive index for tumor cell density and proliferation. This ratio can be used as a marker of tumor infiltration. Cho/NAA reaches the highest values in anaplastic astrocytomas and GBM. Elevated Cho/NAA values can be found in the solid components of the mass when the tissue is still well perfused and oxygenated. Once components of the mass become hypoxic or their apoptotic index increases, a significant drop of the Cho signal occurs in those areas, sometime associated with accumulation of lipids. The Cr signal also changes during this malignant transformation. Cr signal is usually normal or slightly elevated in differentiated and oxygenated astrocytomas. Elevation of Cr is more commonly seen in astrocytoma infiltrating the cortex compared to those growing in the white matter. In astrocytomas infiltrating the gray matter in the cortex or in the basal ganglia and thalami, both Cho and Cr may be elevated with the Cho/Cr approaching unity or even slightly below. Then Cr may drop significantly when new clones with greater proliferation and less differentiating capacities will arise and prevail. Despite several 1H-MRSI studies which have reported high diagnostic accuracy in glioma grading, the possibility that 1H-MRS may
soon replace tissue diagnosis and grading is remote. Only one  study, however, has compared accuracy of 1H-MRSI with contrast-enhanced MR imaging, which is considered the reference standard. Multivoxel studies have shown consistently that there is a trend with HGG having higher Cho/NAA and Cho/Cr compared with LGG. This is especially true if the voxel with the maximum Cho/NAA ratio is used for the analysis. However, 95% confidence intervals are wide and substantial overlapping of results between WHO grade II and III and between grades III and IV makes assignment of grade difficult on the individual patient. This is confirmed by ranking all gliomas by the Cho/NAA or Cho/Cr ratio, respectively. There is a continuum from slow-growing to fast-growing gliomas, with the LGG more frequently represented on the left side of the plot, the WHO III in the middle and the HGG on the right side. The introduction of cut-offs for Cho/NAA and Cho/Cr has been unsuccessful because of too much overlap between different grades. In addition, different medical centers typically use different acquisition protocols and analysis techniques, and threshold values (if they exist) may be difficult to compare between sites. A further complicating factor is metabolic heterogeneity within lesions – for tumor grading purposes, it is unclear exactly where to measure the spectrum from. Logically, it would appear that the use of MRSI to search for the most abnormal voxel (e.g. highest Cho) within the lesion should find the part of the lesion with the highest density of tumor cells. In many cases, the center of the lesion may be necrotic and therefore not a good location to measure the spectrum from.
In conclusion, metabolite abnormalities in gliomas are distributed along a spectrum from grade II through grade IV. However, higher Cho/NAA and Cho/Cr ratios generally suggest a faster growing neoplasm and higher grade neoplasm.
Can MRS predict patient survival?
The long-standing success of the cytogenetic classification proposed by Bailey and Cushing in 1926 is due to a strong correlation between histopathological criteria for grading and prognosis. Whether MRS parameters can predict survival rate has only been the focus of limited studies – one recent study found that high creatine levels in grade II gliomas were a predictor of malignant transformation and decreased survival time. Other studies in pediatric brain tumors have also found high Cho levels (e.g. normalized Cho, or ratio of Cho/NAA) to be prognostic of poor outcome and decreased survival time.
Is it an oligodendroglioma?
The fifth and last FAQ is a fascinating question, because it is also the fundamental question patients often ask at presentation: is this an indolent tumor that will eventually respond to therapy? Diagnosis of oligodendroglioma with histopathology is challenging since even the most experienced neuropathologists have not yet agreed on diagnostic criteria. An oligodendroglioma in Paris may be called a diffuse astrocytoma in Baltimore or an oligoastrocytoma in Milan. Oligodendroglioma is a neoplasm with higher cellular density, vascular density, mitotic and apoptotic indices than a diffuse astrocytoma. Notwithstanding, the prognosis is better for oligodendrogliomas than their astrocytoma “cousins.” The oligodendroglioma is ambiguous also on MR spectroscopy: the Cho peak may be very highly elevated, Cr may be absent, and lactate present – all features of a more malignant neoplasm, yet the prognosis is generally more favorable compared to an astrocytoma. Oligodendrogliomas are “mischievous” also on MR perfusion: rCBV may be very elevated because of high capillary density despite low level of angiogenesis. On post-contrast MRI, oligodendroglioma may sometimes be enhanced as well, although this is not typically seen in low-grade oligodendrogliomas. LOH on chromosome arms 1p and 19q is an early genetic event in the histogenesis of oligodendroglioma and it is considered a genetic hallmark: this feature is present in up to 90% of oligodendrogliomas and 50% of oligoastrocytomas.
Diagnostic accuracy of MRS
In general, multivoxel 1H-MRSI studies using linear discriminant analysis have shown higher diagnostic accuracy than single-voxel 1H-MRS studies. In one large MRSI study, Preul et al. showed that it correctly classified 104 of 105 patients with 4 tumor histotypes (LG astrocytoma, HGG, meningioma, and metastasis). Tate et al., with single-voxel 1H-MRS and linear discriminant analysis, correctly classified 133 of 144 patients in 3 rather broad tumor types (LG astrocytoma, HG tumors [AA, GBM and metastasis] and meningioma. Herminghaus et al., with single-voxel 1H-MRS and linear discriminant analysis in 94 consecutive gliomas, reported a success rate of 86% in grading glioma and a 95% success rate in differentiating LGG from HGG. Recently, in a retrospective 1H-MRSI study on 69 patients, Hourani et al. have shown a high rate of success (84%) for correctly classifying brain tumors from other non-neoplastic brain lesions (i.e. stroke, demyelination, stable undiagnosed lesions). When MR perfusion was added to the analysis, diagnostic accuracy was unchanged with a sensitivity of 72% and specificity of 92%. 1H-MRSI was also shown to be more accurate than MR perfusion in a subset of patients in the same study. However, it is fair to note that this study showed that both 1H-MRSI and MR perfusion techniques may misclassify patients with LGG as non-neoplastic lesions, and vice versa.
The accuracy of 1H-MRSI, integrated with other MR imaging methods, in diagnosing intraxial focal brain masses with a strategically designed algorithm was determined in a study with 111 patients. Diagnosis for each patient was made after collecting imaging data with contrast-enhanced MRI, diffusion MR, perfusion MR, and 1H-MRSI in this order. Accuracy, sensitivity, and specificity of the strategy, respectively, were 90%, 97%, and 67% for discrimination of neoplastic from non-neoplastic processes, 90%, 88%, and 100% for discrimination of HG from LG neoplasms, and 85%, 84%, and 87% for discrimination of HG neoplasms and lymphoma from LG neoplasms and non-neoplastic diseases. The algorithm was strategically built with eight nodes which required imaging input.
MRS imaging-guided therapy planning and monitoring
When the diagnosis of glioma is suspected, surgery is usually the treatment of choice. When the mass is located in the dominant hemisphere and in particular near or within eloquent areas, the risk of postoperative sequelae is higher. However, in the majority of patients, surgery is performed to obtain a pathological specimen so as to establish a definitive diagnosis, to offer a prognosis and to evaluate possible additional or alternative treatment. 1H-MRS has been proposed as a valuable presurgical planning tool to identify the most aggressive components within the tumor volume. Integration of segmented Cho/NAA maps fused with three-dimensional T1-weighted MR images in a neuronavigational system has been accomplished. These hybrid images have been used for frameless stereotaxy and MR spectroscopy-guided biopsy sampling. A correlation study of 1H-MRSI with histopathology in 76 biopsy specimens has found a negative linear correlation (r = –0.905, P < 0.001) of NAA concentration and a positive exponential correlation for Cho (r = 0.769, P < 0.001) and Cho/NAA (r = 0.885, P < 0.001) with increasing tumor infiltration (indicated by tumor cell nuclei/whole cell nuclei on histopathology).
In patients diagnosed with WHO II glioma who had a total tumor resection, there is no indication for additional therapy. Chemotherapy with procarbazine, CCNU and vincristine (PCV) may be indicated in patients older than 40 years, with a large residual tumor volume after surgery, or when a diagnosis of oligodendroglioma, MOA, or gemistocytic astrocytoma is made. As already mentioned, these tumors will respond well to chemotherapy, especially if molecular genetics shows LOH on chromosome 1p and 19q. Chemotherapy with PVC is mandatory in anaplastic oligodendroglioma and MOA (WHO III). Additional treatment with radiotherapy and chemotherapy with temozolomide (TMZ) is administered in patients diagnosed with anaplastic astrocytoma (WHO III) and GBM (WHO IV). The benefit of a second surgery at recurrence is uncertain, and new clinical trials are needed to assess its effectiveness. Upon tumor recurrence, chemotherapy with TMZ or PVC may improve survival and it is a reasonable option.
Spectroscopic imaging is valuable to target volumes for radiotherapy and to evaluate response to therapy. Incorporation of 1H-MRSI into the treatment planning process may have the potential to improve control while minimizing side effects and complications. 1H-MRSI may be useful in monitoring therapeutic response in patients with brain tumors, as it was demonstrated in a longitudinal study in a woman with non-Hodgkin lymphoma who had a complete and persistent favorable response to radiotherapy. 
In following patients post-radiation therapy, the differential diagnosis between recurrent tumor and delayed radiation necrosis (DRN) is one common dilemma. 1H-MRSI has been shown to be useful to improve diagnostic acumen, while conventional MR imaging often cannot differentiate the two entities. Detection of an abnormally elevated Cho signal suggests the diagnosis of recurrent tumor. Alternatively, if an elevated Cho peak is not found within areas of T2-signal abnormality or contrast enhancement, the diagnosis of DRN (or predominantly radiation necrosis) would be suggested. In one of the first applications of 1HMRSI as a presurgical planning tool, Cho elevation was accurate in depicting areas of recurrent tumor within lesions that may be confused with DRN on conventional MR imaging. The detection of lipid signals is not a discriminant sign, despite lipid being usually found in a larger amount in areas of DRN than in areas of recurrent tumor. Lactate may be found in recurrent tumors, but is not a discriminant factor, either.
Special considerations about pediatric brain tumors
Brain tumors are the second most common group of neoplasms in childhood, following leukemia. Among the primary neuroepithelial brain neoplasms, the percentage of LG neoplasms is much higher in children than in adults. In contrast, metastatic brain tumors are rare in children, while they represent 30% of neoplasms in adults. Age at onset and location are two very important diagnostic and prognostic factors. Low-grade astrocytomas (pilocytic or diffuse astrocytoma, pleomorphic xanthoastrocytoma, and subependymal giant cell astrocytoma – usually in association with tuberous sclerosis) are the most frequent (about 35–40%), followed by primitive neuroectodermal tumors (PNET) or medulloblastoma (about 20%) and ependymoma (about 10–15%). Craniopharyngioma, HGG, ganglioglioma, and germ cell tumors are less common primary brain tumors in children. Pilocytic astrocytoma is the most common tumor in children and may occur in cerebellum, hypothalamus, and optic nerves. Diffuse astrocytomas are relatively frequently located in the brain stem and, in particular, in the pons, where they are usually poorly defined and cause diffuse enlargement. PNET is the second most frequent and it occurs in the cerebellar vermis during the first decade of life. Supratentorial locations are rare. These tumors are composed of densely packed cells with hyperchromatic nuclei and scant cytoplasm. Focal areas of hemorrhage and necrosis are frequently found. The presence of leptomeningeal metastases is often associated with PNET, therefore staging with MRI of the spine must be performed before surgery.
Ependymoma is relatively more common in childhood than in adults: 65% occur in the posterior fossa, 25% supratentorially, and 10% in the spinal cord. Histologically, they are very well circumscribed and separated from the brain. Ependymoma may show features of anaplasia with high mitotic rate, cellular pleomorphism, and necrosis in about 25% of cases.
In 1995, Wang et al. proposed the use of single-voxel 1H-MRS at long TE to differentiate 30 pediatric brain tumors occurring in the posterior fossa. It was shown that the Cho/NAA ratio was higher in PNET, while Cho/Cr was significantly higher in ependymomas. More recently, it has been shown that an abnormally elevated taurine signal (3.4 ppm) can be detected with single voxel 1H-MRS at short TE in PNET, but not in other posterior fossa tumors. In this study on 29 children, large standard deviations in Cho, Cr, and myo-inositol were found in PNET and other types of tumors: Cho/NAA was significantly higher in PNET (P < 0.001). In a study on 14 children with hemispheric brain neoplasms, elevated Cho/NAA and Cho/Cr were associated with shorter survival and poor prognosis.
In pediatric patients, as in adults, the Cho/NAA ratio is higher in tumors with higher cell density, associated with a poor prognosis, and this is the most valuable parameter. As in adults, definitive diagnosis of tumor type and grade cannot usually be made based on the spectrum alone because of too much overlap between groups; however, MRS may provide useful, complementary information to assist other imaging modalities in making a differential diagnosis.
Conclusions
MR spectroscopy and MRSI are useful techniques that provide unique metabolic information for characterization of brain tumors in vivo. 1H-MRS is a valuable tool in differentiating active and recurrent tumor from non-neoplastic lesions, and from therapy-related tissue injury. 1H-MRS also provides an estimate of cell tumor density and proliferation, and may detect transformation of tumor to a more malignant subtype. Several technical, logistical, and financial issues somewhat inhibit the widespread application of this methodology, including the lack of widely accepted, standardized acquisition and analysis protocols that could be used in multicenter studies. Despite these challenges, 1H-MRS will continue to be an important tool for physicians working to improve brain tumor diagnosis, prognosis, and therapy.

What’s New

 
August/14/2007
Inomed ISIS Intraoperative neurophysiological monitoring started to function in all our related surgeries.
Oct /07/2009
The author celebrating 30 years experience in neurosurgery.
Nov /27/2013
Magnetom Skyra 3 tesla with all clinical applications is running and intraoperative MRI monitoring started.
  Copyright [2014] [CNS Clinic-JORDAN]. All rights reserved